Evaluate Returns: Part 1





Kerry Back

Overview

  • Suppose we’ve run a backtest and saved the portfolio returns.
  • Evaluate raw returns: accumulation, drawdowns, mean, std dev, Sharpe ratio
  • Evaluate relative to benchmarks
    • Compare to the market on a beta-adjusted basis
    • Compare to the market and other factors (attribution analysis)

Example returns

  • Best 200 stocks each month, equally weighted
  • Worst 200 stocks each month, equally weighted
  • 150/50 portfolio = 1.5 \(\times\) best - 0.5 \(\times\) worst
  • Market return from French’s data library

Accumulation plots

  • (1+ret).cumprod()
  • Also on log scale

Drawdowns

  • A drawdown occurs whenever your portfolio value is not as high as it once was
  • The percent drawdown is (current value - prior max) / prior max.

Statistics

  • mean and std dev of monthly returns
  • annualize mean by \(\times\) 12
  • annualize std dev by \(\times\) sqrt(12)
  • compute monthly excess returns (in excess of T-bill rate), mean and std dev and Sharpe ratio = mean / std dev
  • annualize Sharpe ratio by \(\times\) sqrt(12)